Vehicle positioning method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a vehicle positioning method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring sensing detection information and odometer data of a vehicle in an automatic parking process; according to the perception detection information, determining point cloud data with boundary semantic element information perceived by the vehicle; performing track plotting processing on the vehicle according to the odometer data to obtain vehicle track plotting information corresponding to the vehicle; performing two-dimensional GICP matching on the vehicle track plotting information and the point cloud data to obtain a GICP matching result; and under the condition that the GICP matching result indicates that the vehicle state does not reach the expected state, carrying out filtering positioning processing on the vehicle track plotting information based on a particle filtering positioning algorithm to obtain the vehicle position information of the vehicle. According to the embodiment of the invention, intelligent parking of the vehicle in a complex environment can be more efficient and reliable.

Description

Technical Field

[0001] The present application relates to the technical field of vehicle positioning, and in particular to a vehicle positioning method, device, electronic equipment and storage medium. Background Art

[0002] With the rapid development of autonomous driving technology, valet memory parking has gradually become an important part of the assisted driving system. In this process, the environment-based memory mapping function is particularly important, which can improve the convenience and safety of user parking.

[0003] The currently commonly used automatic parking technology uses the image information captured by the camera to extract, match and build maps through the visual SLAM (Simultaneous Localization and Mapping) algorithm, thereby achieving vehicle positioning and navigation. However, the above method is sensitive to lighting conditions and is easily affected by factors such as shadows and reflections. In an environment with a lack of texture, feature point extraction may become difficult, resulting in a decrease in positioning accuracy, thereby reducing the reliability and efficiency of automatic parking. Summary of the invention

[0004] The embodiments of the present application provide a vehicle positioning method, device, electronic device and storage medium to solve the problem that the method of using SLAM algorithm to realize vehicle positioning and navigation in related technologies is sensitive to lighting conditions and easily affected by factors such as shadows and reflections. In an environment lacking texture, feature point extraction may become difficult, resulting in a decrease in positioning accuracy, thereby reducing the reliability and efficiency of automatic parking.

[0005] In order to solve the above technical problems, the embodiments of the present application are implemented as follows:

[0006] In a first aspect, an embodiment of the present application provides a vehicle positioning method, the method comprising:

[0007] Obtaining the vehicle's perception detection information and odometer data during the automatic parking process;

[0008] Determining point cloud data with boundary semantic element information perceived by the vehicle according to the perception detection information;

[0009] Performing dead reckoning processing on the vehicle according to the odometer data to obtain vehicle dead reckoning information corresponding to the vehicle;

[0010] Performing two-dimensional GICP matching on the vehicle dead reckoning information and the point cloud data to obtain a GICP matching result;

[0011] When the GICP matching result indicates that the vehicle state has not reached the expected state, the vehicle dead reckoning information is filtered and positioned based on a particle filter positioning algorithm to obtain the vehicle position information of the vehicle.

[0012] In a second aspect, an embodiment of the present application provides a vehicle positioning device, the device comprising:

[0013] A vehicle data acquisition module, used to acquire the vehicle's perception detection information and odometer data during the automatic parking process;

[0014] A point cloud data determination module, used to determine the point cloud data with boundary semantic element information perceived by the vehicle according to the perception detection information;

[0015] A dead reckoning information acquisition module, configured to perform dead reckoning processing on the vehicle according to the odometer data to obtain vehicle dead reckoning information corresponding to the vehicle;

[0016] A GICP matching result acquisition module, used for performing two-dimensional GICP matching on the vehicle dead reckoning information and the point cloud data to obtain a GICP matching result;

[0017] The vehicle position information acquisition module is used to filter and locate the vehicle track calculation information based on the particle filter positioning algorithm to obtain the vehicle position information of the vehicle when the GICP matching result indicates that the vehicle state has not reached the expected state.

[0018] In a third aspect, an embodiment of the present application provides an electronic device, including:

[0019] A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program implements any of the above-mentioned vehicle positioning methods when executed by the processor.

[0020] In a fourth aspect, an embodiment of the present application provides a readable storage medium, which, when instructions in the storage medium are executed by a processor of an electronic device, enables the electronic device to execute any of the vehicle positioning methods described above.

[0021] In an embodiment of the present application, the perception detection information and odometer data of the vehicle during the automatic parking process are obtained. Based on the perception detection information, the point cloud data with boundary semantic element information perceived by the vehicle is determined. The vehicle is tracked according to the odometer data to obtain the vehicle track information corresponding to the vehicle. Two-dimensional GICP matching is performed on the vehicle track information and the point cloud data to obtain the GICP matching result. When the GICP matching result indicates that the vehicle state has not reached the expected state, the vehicle track information is filtered and positioned based on the particle filter positioning algorithm to obtain the vehicle position information of the vehicle. The embodiment of the present application completes the semantic GICP matching and particle filter positioning algorithm based on the semantic element information to realize the parking positioning of the vehicle, so as to perform real-time identification and analysis of the surrounding environment, and can effectively complete the positioning and repositioning functions through environmental information, which can make the intelligent parking of the vehicle in a complex environment more efficient and reliable.

[0022] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0024] Figure 1 A flowchart of a vehicle positioning method provided in an embodiment of the present application;

[0025] Figure 2 A schematic diagram of a vehicle positioning process provided in an embodiment of the present application;

[0026] Figure 3 A schematic diagram of the structure of a vehicle positioning device provided in an embodiment of the present application;

[0027] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0029] Reference Figure 1 , shows a flow chart of the steps of a vehicle positioning method provided by an embodiment of the present application, such as Figure 1 As shown, the vehicle positioning method may include: step 101, step 102, step 103, step 104 and step 105.

[0030] Step 101: Acquire the perception detection information and odometer data of the vehicle during the automatic parking process.

[0031] The embodiments of the present application can be applied to scenarios where vehicle parking positioning is performed by combining GICP matching and particle filter positioning algorithms.

[0032] Perception detection information refers to data about the surrounding environment collected and processed in real time by various sensors (such as cameras, ultrasonic radars, laser radars, etc.) carried by the vehicle during the automatic parking process. In this example, the perception detection information may include: lane edge information, parking space information, lane marking information, zebra crossings, speed bumps, pillars, and other information.

[0033] Odometer data refers to the vehicle's distance, speed, acceleration, direction and other motion parameters measured and recorded in real time by the vehicle's internal odometer (usually composed of wheel encoders, inertial measurement units, etc.) during driving. These data are crucial for vehicle position estimation, path planning and motion control. During automatic parking, odometer data can help the vehicle accurately track its driving trajectory and ensure the accuracy and safety of parking actions.

[0034] During the process of automatic parking, the vehicle's perception detection information and odometer data can be obtained. Specifically, the vehicle's sensors can scan the surrounding environment in real time, and transmit the collected data to the central processor through a data line for processing. The central processor will analyze and fuse these data to generate the perception detection information required by the vehicle. At the same time, the vehicle's odometer system will measure the rotation of the wheels through the wheel encoder, and measure the odometer data such as the vehicle's acceleration and angle change through the inertial measurement unit. These data will be recorded in real time and transmitted to the vehicle's control system for subsequent path planning and motion control.

[0035] After acquiring the perception detection information and odometer data of the vehicle during the automatic parking process, step 102 is executed.

[0036] Step 102: Determine the point cloud data with boundary semantic element information perceived by the vehicle based on the perception detection information.

[0037] After obtaining the perception detection information of the vehicle during the automatic parking process, the point cloud data with boundary semantic element information perceived by the vehicle can be determined based on the perception detection information. Specifically, the sensors carried by the vehicle will scan the surrounding environment in real time, and transmit the collected data to the central processor through the data line for processing. The central processor will analyze and fuse the data to generate the perception detection information required by the vehicle. Then, the perception detection information can be discretized to obtain point cloud data. The implementation process can be described in detail in combination with the following specific implementation method.

[0038] In a specific implementation of the present application, the above step 102 may include:

[0039] Sub-step A1: Determine the position information of the perception detection information relative to the vehicle based on the conversion relationship between the inverse perspective mapping image corresponding to the perception detection information and the real position.

[0040] In this embodiment, Inverse Perspective Mapping (IPM) is a commonly used technology that can convert two-dimensional information in an image into three-dimensional information that is closer to the real world, thereby helping the vehicle to understand its surrounding environment more accurately.

[0041] The real position refers to the location of the vehicle and its surroundings in three-dimensional physical space.

[0042] After obtaining the perception detection information, the position information of the perception detection information relative to the vehicle can be determined based on the conversion relationship between the inverse perspective mapping image corresponding to the perception detection information and the real position. Specifically, a mapping relationship from the image plane to the ground plane can be constructed. In the image after inverse perspective mapping, features related to vehicle parking, such as lane lines, parking space boundaries, obstacle edges, etc., are extracted, and the extracted features are matched with known environmental maps or pre-defined parking space features. Combined with the matching results and the vehicle's odometer data (such as information provided by wheel speed sensors, IMU, etc.), an optimization algorithm (such as least squares method, Kalman filtering, particle filtering, etc.) is used to accurately determine the position of the perception detection information relative to the vehicle.

[0043] After determining the position information of the perception detection information relative to the vehicle based on the conversion relationship between the inverse perspective mapping image corresponding to the perception detection information and the real position, sub-step A2 is executed.

[0044] Sub-step A2: converting the perception detection information into a global coordinate system corresponding to the vehicle according to the posture information to obtain conversion coordinates corresponding to the perception detection information.

[0045] After determining the position information of the perception detection information relative to the vehicle according to the conversion relationship between the inverse perspective mapping image corresponding to the perception detection information and the real position, the perception detection information can be converted to the global coordinate system corresponding to the vehicle according to the position information to obtain the conversion coordinates corresponding to the perception detection information. Specifically, the global coordinate system of the vehicle is usually established with the vehicle as the center, wherein the initial position of the vehicle is the origin, the driving direction of the vehicle is the positive direction of the X axis, the left side of the vehicle is the positive direction of the Y axis, and the direction perpendicular to the ground is the positive direction of the Z axis. Then, using the known position information (position, attitude), each point in the perception detection information is converted from the original coordinate system to the global coordinate system, which involves rotation and translation transformation, wherein the rotation transformation is used to adjust the direction of the point from the camera coordinate system or the image coordinate system to the direction in the global coordinate system, and the translation transformation is used to move the point from the origin of the camera or the image to the origin of the global coordinate system. By applying the formula of rotation and translation transformation, the coordinates of each point in the global coordinate system are calculated, and these conversion coordinates can now be combined with the global path planning, obstacle detection, navigation and other systems of the vehicle to achieve more accurate and reliable autonomous driving functions.

[0046] After converting the perception detection information to the global coordinate system corresponding to the vehicle according to the posture information and obtaining the conversion coordinates corresponding to the perception detection information, execute sub-step A3.

[0047] Sub-step A3: Discretely process the perception detection information according to the conversion coordinates to obtain point cloud data with boundary semantic element information perceived by the vehicle.

[0048] After the perception detection information is converted to the global coordinate system corresponding to the vehicle according to the posture information and the conversion coordinates corresponding to the perception detection information are obtained, the perception detection information can be discretized according to the conversion coordinates to obtain the point cloud data with boundary semantic element information perceived by the vehicle. Specifically, discretizing the conversion coordinates means converting continuous spatial data (such as the point cloud obtained by the laser radar scan) into a series of discrete points or pixels. For the laser radar data, each scanning point is a three-dimensional coordinate point; for the camera image, it may be necessary to divide the image into multiple pixel regions and extract the features of each region. After the discretization process, the next step is to extract information with boundary semantic elements, which involves using image processing or point cloud processing algorithms to identify different categories of objects (such as vehicles, pedestrians, trees, etc.) and their boundaries. After the boundary semantic elements are extracted, point cloud data with this information can be generated, and then the coordinates of each point can be associated with its category and boundary information. In this way, each point not only contains its position information in the global coordinate system, but also contains semantic information about the category of the object or region to which it belongs and boundary information.

[0049] Step 103: performing dead reckoning processing on the vehicle according to the odometer data to obtain vehicle dead reckoning information corresponding to the vehicle.

[0050] After obtaining the odometer data of the vehicle, dead reckoning processing may be performed on the vehicle based on the odometer data to obtain vehicle dead reckoning information corresponding to the vehicle.

[0051] In a specific implementation, dead reckoning (also known as dead reckoning or track prediction) is a method of estimating the current position based on a known starting position and a series of incremental motion measurements. In vehicle navigation, this typically involves using odometer data to track the movement of the vehicle. Specifically, the dead reckoning process can be: first set the initial position of the vehicle, which is usually known (for example, by setting a positioning method), and then use the odometer data to update the estimated position of the vehicle, which involves adding the distance and direction changes of each time step to the position estimate of the previous time step. For motion on a two-dimensional plane, simple vector addition can be used to update the position. For motion in three-dimensional space, more complexities need to be considered, such as attitude changes such as pitch and roll.

[0052] Step 104: Perform two-dimensional GICP matching on the vehicle dead reckoning information and the point cloud data to obtain a GICP matching result.

[0053] GICP (Generalized Iterative Closest Point) is a method for point cloud registration. GICP matching is a 3D point cloud registration algorithm that integrates Point-To-Point and Point-To-Plane. It introduces probability information and covariance matrix to form a unified registration model. In GICP matching, it is assumed that there are two matched point sets A and B, and each point obeys a Gaussian distribution. The correlation between two corresponding points is estimated by calculating the covariance matrix between them, and then the weighted least squares method is used to optimize the rigid body transformation matrix to minimize the distance between the point clouds.

[0054] After obtaining the vehicle dead reckoning information and point cloud data, two-dimensional GICP matching can be performed on the vehicle dead reckoning information and the point cloud data to obtain a GICP matching result. Specifically, the point cloud data can be classified according to the boundary semantic element information, an initial rigid body transformation matrix can be constructed according to the vehicle dead reckoning information, and then two-dimensional GICP matching can be performed. The implementation process can be described in detail in combination with the following specific implementation method.

[0055] In a specific implementation of the present application, the above step 104 may include:

[0056] Sub-step B1: classifying the point cloud data according to the boundary semantic element information to obtain point cloud data of different categories.

[0057] In this embodiment, after obtaining point cloud data with boundary semantic element information, the point cloud data can be classified according to the boundary semantic element information to obtain point cloud data of different categories, that is, point cloud data classification. The point cloud data can be classified by different categories of point cloud boundary points defined by the boundary semantic element information to obtain point cloud data of different categories.

[0058] Sub-step B2: constructing an initial rigid body transformation matrix based on the vehicle dead reckoning information.

[0059] The initial rigid body transformation matrix can be used to describe the relative posture and position relationship between the vehicle at different times or different positions.

[0060] After obtaining the vehicle dead reckoning information, the initial rigid body transformation matrix can be constructed based on the vehicle dead reckoning information. The specific process can be:

[0061] 1. Calculate the relative position and attitude change of the vehicle from the reference coordinate system to the target coordinate system based on the vehicle dead reckoning data (such as GPS position, IMU attitude information, etc.). The reference coordinate system is the coordinate system corresponding to the position or attitude of the vehicle or sensor when the dead reckoning begins, and the target coordinate system is the coordinate system corresponding to the position or attitude that the vehicle or sensor needs to reach or align with during the dead reckoning process.

[0062] 2. According to the calculated attitude changes (such as pitch angle, yaw angle, roll angle, etc.), a rotation matrix is ​​constructed using a rotation matrix construction method (such as the Rodriguez rotation formula or the product representation of the rotation matrix).

[0063] 3. Construct a translation vector based on the calculated relative position change (such as the increment of X, Y, and Z coordinates).

[0064] 4. Combine the rotation matrix and translation vector into a 4x4 rigid body transformation matrix.

[0065] Sub-step B3: Perform two-dimensional GICP matching on the initial rigid body transformation matrix and the point cloud data of different categories to obtain the GICP matching result.

[0066] Finally, two-dimensional GICP matching can be performed on the initial rigid body transformation matrix and point cloud data of different categories to obtain GICP matching results. The GICP matching process can be described in detail in conjunction with the following specific implementation method.

[0067] In another specific implementation of the present application, the above sub-step B3 may include:

[0068] Sub-step C1: converting the point cloud data of different categories into two-dimensional point cloud data on the two-dimensional plane where the vehicle is located; the two-dimensional point cloud data includes: source point cloud data of the vehicle and target point cloud data of the vehicle's perceived environment.

[0069] In this embodiment, after obtaining different categories of point cloud data, the different categories of point cloud data can be converted into two-dimensional point cloud data on the two-dimensional plane where the vehicle is located. The two-dimensional point cloud data may include: source point cloud data of the vehicle and target point cloud data of the vehicle's perceived environment.

[0070] Sub-step C2: During each iteration, a nearest neighbor search algorithm is used to find the nearest point of each point in the source point cloud data from the target point cloud data to form a corresponding point pair set.

[0071] Nearest neighbor search, also known as nearest neighbor search (NNS) or closest point search, refers to the process of finding the element with the smallest distance to a given query item within a certain distance metric and a search space. In this example, the nearest neighbor search algorithm can be but is not limited to the KD-tree search algorithm. KD-tree (K-dimension tree) is a data structure that divides data points in k-dimensional space. Using KD-tree can greatly reduce the time complexity of nearest neighbor search. After constructing the KD-tree, the point closest to the target point can be found by recursively searching down the binary tree and backtracking the search path when necessary.

[0072] In each iteration, algorithms such as KD-Tree can be used to find the nearest point of each point in the source point cloud data from the target point cloud data, so that a set of corresponding point pairs can be formed.

[0073] In each iteration, after using the nearest neighbor search algorithm to find the nearest point of each point in the source point cloud data from the target point cloud data to form a corresponding point pair set, sub-step C3 is executed.

[0074] Sub-step C3: For each corresponding point, the Euclidean distance between the source point and the target point after being transformed by the initial rigid body transformation matrix is ​​calculated to obtain a residual.

[0075] For each corresponding point, the Euclidean distance between the source point (i.e., the point in the source point cloud data) and the target point (i.e., the point in the target point cloud data) after being transformed by the current rigid body transformation matrix can be calculated to obtain the residual.

[0076] Sub-step C4: processing the rotation parameters and translation parameters of the vehicle based on the residual to construct a residual Jacobian equation.

[0077] Then, based on the residual, the partial derivatives of the rotation parameters and translation parameters of the vehicle can be calculated to construct a residual Jacobian matrix, which describes the effect of parameter changes on the residual. In this embodiment, the defined derivative mapping tangent vector space coincides with the plane rotation manifold (a set of all rotation operations on a two-dimensional plane) and SO(2), that is, the plane rotation manifold describes all rotation operations on a two-dimensional plane, and the SO(2) group represents these rotation operations through a rotation matrix. Therefore, the residual Jacobian equation in this embodiment only considers the two-dimensional plane where the vehicle is located.

[0078] The formula for the residual Jacobi equation can be shown as follows:

[0079]

[0080] In the above formula (1), is the residual Jacobi equation, D is the derivative function, R is the two-dimensional rotation parameter of the vehicle, v is the translation parameter of the vehicle, is a 2*2 Lie group space, representing the set of real numbers.

[0081] Sub-step C5: iteratively process the residual Jacobian equation to obtain the GICP matching result.

[0082] Finally, the residual Jacobian equation can be iterated using an optimization algorithm (such as the Gauss-Newton method, the Levenberg-Marquardt method, etc.) to minimize the residual and update the rigid body transformation matrix. After each iteration, check whether the convergence conditions are met. Convergence conditions may include the amount of residual reduction being less than a certain threshold, the number of iterations reaching an upper limit, etc. During the iteration process, the inliers (i.e., points that meet the model assumptions) can also be screened according to the size of the residuals, and only the inliers are used to update the rigid body transformation matrix. This helps to improve the accuracy and robustness of the matching. The output GICP matching results may include: the optimized rigid body transformation matrix, inliers, whether convergence, and residual values.

[0083] Among them, the optimized rigid body transformation matrix is ​​one of the main outputs of the GICP algorithm, which represents the optimal rigid body transformation of the source point cloud relative to the target point cloud.

[0084] The inliers are a set of screened inliers, which are considered as valid supports for the alignment results during the iteration process.

[0085] Whether converged is a Boolean value, indicating whether the GICP algorithm successfully converged to a solution.

[0086] The residual value represents the final residual value or the sum of residuals, which is used to evaluate the accuracy of the alignment result.

[0087] After performing two-dimensional GICP matching on the vehicle dead reckoning information and the point cloud data to obtain the GICP matching result, step 105 is executed.

[0088] Step 105: When the GICP matching result indicates that the vehicle state has not reached the expected state, filtering and positioning processing is performed on the vehicle dead reckoning information based on a particle filter positioning algorithm to obtain the vehicle position information of the vehicle.

[0089] After performing two-dimensional GICP matching on the vehicle dead reckoning information and point cloud data to obtain the GICP matching result, it is possible to judge whether the vehicle state has reached the expected state (i.e., whether the vehicle state has reached a stable state) based on the GICP matching result. Specifically, it is possible to judge whether the overall state has reached a stable state based on whether the GICP converges and the proportion of internal points to the total number of points. If it has, it will be further judged whether the residual value is within a predetermined range. If it is within the predetermined range, it is considered that the entire state is stable and well-converged. At this time, the GICP result can be used directly and passed into the fusion module. The implementation process can be described in detail in conjunction with the following specific implementation method.

[0090] In a specific implementation of the present application, after the above step 104, the following may also be included:

[0091] Step D1: When the GICP matching result indicates that the vehicle state has reached the expected state, the point cloud data is transformed to a position aligned with the vehicle dead reckoning information using the target rigid body transformation matrix to obtain transformed point cloud data.

[0092] In this embodiment, when the GICP matching result indicates that the vehicle state has reached the expected state, the target rigid body transformation matrix (i.e., the optimized rigid body transformation matrix) can be used to transform the point cloud data to a position aligned with the vehicle dead reckoning information to obtain transformed point cloud data. Specifically, the coordinates of each point cloud can be multiplied by the target rigid body transformation matrix through matrix multiplication to obtain new point cloud data, i.e., transformed point cloud data.

[0093] Step D2: Fusing the transformed point cloud data with the vehicle dead reckoning information to obtain the vehicle positioning position of the vehicle.

[0094] After obtaining the transformed point cloud data, the transformed point cloud data and the vehicle dead reckoning information can be fused to obtain the vehicle positioning position of the vehicle. Specifically, it is first necessary to ensure that the transformed point cloud data and the vehicle dead reckoning information are synchronized in time, that is, they correspond to the data at the same time point. Since the point cloud data and the vehicle dead reckoning information have been aligned through the target rigid body transformation matrix, this step is mainly to verify the accuracy of the alignment. Then, different weights can be assigned to the transformed point cloud data and the vehicle dead reckoning information according to the accuracy and reliability of different data. Finally, the transformed point cloud data and the vehicle dead reckoning information can be weighted summed to obtain the vehicle positioning position of the vehicle.

[0095] When the GICP matching result indicates that the vehicle state reaches the expected state, the particle filter positioning algorithm can be used to filter and position the vehicle track calculation information to obtain the vehicle position information of the vehicle.

[0096] Among them, the particle filter positioning algorithm is a positioning technology based on the Monte Carlo method, which uses a set of particles randomly extracted from the posterior probability to approximate the target probability density function, thereby achieving positioning. The particle filter positioning process can be described in detail in combination with the following specific implementation method.

[0097] In a specific implementation of the present application, the above step 105 may include:

[0098] Sub-step E1: In the initialization stage of the particle filter, the particle filter range is determined according to the GICP matching result.

[0099] In this embodiment, in the initialization stage of the particle filter, the particle filter range can be determined according to the GICP matching result. Specifically, after the GICP algorithm obtains the transformation relationship between the two point clouds, the approximate position and posture of the vehicle can be calculated. Then, with this position and posture as the center, a relatively small search range is set for the initialization of the particle filter. In this way, the particle filter only needs to search and iterate within this smaller range, thereby greatly improving the calculation efficiency and positioning accuracy.

[0100] Sub-step E2: within the particle filtering range, based on the motion model, the particle state of the vehicle at the previous moment and the vehicle dead reckoning information, a state prediction is performed on each particle to generate a new particle state, wherein each particle represents a point cloud data in the vehicle dead reckoning information, and the particle state refers to the position state of each point cloud data in the vehicle dead reckoning information.

[0101] Then, within the particle filter range, the state of each particle can be predicted based on the motion model and the particle state of the vehicle at the last moment and the vehicle dead reckoning information to generate a new particle state, wherein each particle represents a point cloud data in the vehicle dead reckoning information, and the particle state refers to the position state of each point cloud data in the vehicle dead reckoning information. That is, within the determined particle filter range, the state of each particle is predicted based on the motion model and the particle state at the last moment (i.e., the position state of the point cloud data in the vehicle dead reckoning information) to generate a new particle state. Here, each particle represents a point cloud data point in the vehicle dead reckoning, and its state describes the position of the point cloud data in space.

[0102] In the specific implementation, the particle filter predicts the state of each particle based on the motion model, the particle state at the previous moment and the control input (such as speed, steering, etc.), and generates a new particle state. This process takes into account the noise of the system to simulate the randomness of actual motion.

[0103] Sub-step E3: after receiving the new vehicle dead reckoning information, calculating the matching degree between the new particle state of each particle and the new vehicle dead reckoning information.

[0104] After receiving new vehicle dead reckoning information, the degree of match between the new particle state of each particle and the new vehicle dead reckoning information can be calculated. That is, when new sensory detection data is passed in, the measurement model can be used to calculate the degree of match between each particle and the observed data. The degree of match is usually measured by calculating the error or distance between the observed data and the particle state.

[0105] Sub-step E4: updating the particle weight of each particle according to the matching degree to obtain an updated weight of each particle.

[0106] After calculating the degree of match between the new particle state of each particle and the new vehicle dead reckoning information, the particle weight of each particle can be updated according to the degree of match to obtain the updated weight of each particle. The weight update is usually based on the likelihood function, which reflects the similarity between the particle state and the observed data. The higher the degree of match, the greater the weight of the particle, indicating that the particle is more likely to represent the true state of the system.

[0107] Sub-step E5: resample all particles according to the updated weights to obtain a resampled particle set.

[0108] During the iteration of particle filtering, the weights of some particles may be significantly higher than other particles, resulting in degradation of particle samples (i.e., the diversity of the particle set is reduced). To solve this problem, resampling is required. The resampling process is to select a new particle set based on the updated weights of the particles, retaining particles with high weights while discarding or reducing particles with low weights.

[0109] The purpose of resampling is to ensure the diversity of the particle set and prevent the algorithm from falling into a local optimal solution.

[0110] Sub-step E6: Processing the resampled particle set to obtain an estimated value of the system state of the vehicle.

[0111] After obtaining the resampled particle set, the resampled particle set can be weighted averaged or combined in other forms (such as median, mode, etc.) to obtain the estimated value of the vehicle's system state. The estimated value represents the most likely value of the system state under the current observation data and motion model constraints.

[0112] Sub-step E7: According to the estimated value of the system state, the vehicle position information of the vehicle is located.

[0113] Furthermore, the vehicle position information of the vehicle can be located based on the estimated value of the system state. That is the ultimate goal of the particle filter algorithm in the vehicle positioning system. Specifically, the particle filter confidence can be calculated based on the updated weight. When the particle filter confidence is greater than the confidence threshold, the vehicle position information of the vehicle can be located based on the estimated value of the system state. In the last step of the particle filter, the estimated value of the system state is obtained by weighted averaging or other forms of combination (such as median, maximum weight particle, etc.). At the same time, the confidence of the particle filter can be indirectly obtained based on the weight value calculated by the likelihood function. The higher the likelihood value, the better the match between the particle state and the observed data, and therefore the higher the confidence.

[0114] The confidence of particle filtering can be calculated in many ways, but it is usually related to the weight distribution of particles. For example, the average or maximum value of all particle weights can be calculated as a measure of confidence, etc. Or the entropy of the weight distribution can be calculated. The smaller the entropy, the more concentrated the distribution of the particle set, and therefore the higher the confidence.

[0115] After obtaining the confidence level, it is necessary to determine whether the value meets the conditions for being output as the final value. This usually depends on the system's requirements for confidence or a preset threshold. If the confidence level is higher than the threshold, the system state estimate obtained by the particle filter can be used as the final output.

[0116] If the confidence is less than or equal to the threshold, it indicates that the current detection effect is poor or there are other interference factors, which cause the particle filter to be unable to accurately estimate the system state. At this time, the vehicle position information of the vehicle can be located based on the vehicle track calculation information of the vehicle at the current moment. Specifically, the track calculation is based on the position, speed and direction of the vehicle at the previous moment, and predicts the current moment through the motion model.

[0117] In the embodiment of the present application, the track calculation considers the displacement and rotation of the vehicle in a short period of time, and uses them as the initial values ​​of 2D GICP and particle filter matching, respectively, to provide the initial position information of 2D GICP and particle filter matching during state update optimization. Then, because of concerns about non-convergence and fluctuations in particle filter matching results, the track calculation will be put into the fusion process at the same time. This part uses Kalman filtering, using track calculation as the prediction amount and particle filter results as the observation amount to complete the smoothing of the output trajectory.

[0118] Next, combine Figure 2 The above implementation process is described in detail.

[0119] like Figure 2As shown, during the automatic parking process of the vehicle, the odometer data and point cloud data of the vehicle can be obtained, wherein the point cloud data can be a discrete point cloud obtained after performing IPM conversion and other processing on the vehicle's perception detection information.

[0120] Then, the odometer data can be used to perform dead reckoning, and the dead reckoning information and point cloud data can be matched in 2D GICP to obtain the GICP matching result. At the same time, it is determined whether the initial value needs to be updated for the particle filter, that is, whether the particle scattering range needs to be reduced through GICP. If so, the updated initial value is given to the particle filter positioning algorithm.

[0121] At the same time, it can be determined whether the GICP results converge. If so, the vehicle positioning is performed directly through the multi-sensor fusion method of the dead reckoning information.

[0122] If the GICP result does not converge, a particle filter matching is performed to obtain the estimated value of the vehicle system state and calculate its confidence. If the confidence is high, the estimated value can be used as the final estimated value output to locate the vehicle position. If the confidence is low, it means that the current particle filter positioning effect is not good, and the current position information can be calculated through the track.

[0123] The vehicle positioning method provided in the embodiment of the present application obtains the perception detection information and odometer data of the vehicle during the automatic parking process. According to the perception detection information, the point cloud data with boundary semantic element information perceived by the vehicle is determined. The vehicle is subjected to track reckoning processing according to the odometer data to obtain the vehicle track reckoning information corresponding to the vehicle. The vehicle track reckoning information and the point cloud data are subjected to two-dimensional GICP matching to obtain the GICP matching result. When the GICP matching result indicates that the vehicle state has not reached the expected state, the vehicle track reckoning information is filtered and positioned based on the particle filter positioning algorithm to obtain the vehicle position information of the vehicle. The embodiment of the present application completes the semantic GICP matching and particle filter positioning algorithm based on the semantic element information to realize the parking positioning of the vehicle, so as to perform real-time recognition and analysis of the surrounding environment, and can effectively complete the positioning and repositioning functions through environmental information, so as to make the intelligent parking of the vehicle in a complex environment more efficient and reliable.

[0124] Reference Figure 3 , shows a schematic diagram of the structure of a vehicle positioning device provided by an embodiment of the present application. Figure 3 As shown, the vehicle positioning device 300 may include the following modules:

[0125] The vehicle data acquisition module 310 is used to acquire the perception detection information and odometer data of the vehicle during the automatic parking process;

[0126] A point cloud data determination module 320, configured to determine the point cloud data with boundary semantic element information perceived by the vehicle according to the perception detection information;

[0127] A dead reckoning information acquisition module 330, configured to perform dead reckoning processing on the vehicle according to the odometer data to obtain vehicle dead reckoning information corresponding to the vehicle;

[0128] A GICP matching result acquisition module 340 is used to perform two-dimensional GICP matching on the vehicle dead reckoning information and the point cloud data to obtain a GICP matching result;

[0129] The vehicle position information acquisition module 350 is used to filter and locate the vehicle dead reckoning information based on a particle filter positioning algorithm to obtain the vehicle position information of the vehicle when the GICP matching result indicates that the vehicle state has not reached the expected state.

[0130] Optionally, the point cloud data determination module includes:

[0131] A posture information determination unit, configured to determine the posture information of the perception detection information relative to the vehicle according to a conversion relationship between an inverse perspective mapping image corresponding to the perception detection information and a real position;

[0132] A conversion coordinate acquisition unit, used to convert the perception detection information into a global coordinate system corresponding to the vehicle according to the posture information, and obtain conversion coordinates corresponding to the perception detection information;

[0133] The point cloud data acquisition unit is used to perform discrete processing on the perception detection information according to the conversion coordinates to obtain point cloud data with boundary semantic element information perceived by the vehicle.

[0134] Optionally, the GICP matching result acquisition module includes:

[0135] A point cloud acquisition unit, used to classify the point cloud data according to the boundary semantic element information to obtain point cloud data of different categories;

[0136] An initial matrix construction unit, used to construct an initial rigid body transformation matrix according to the vehicle dead reckoning information;

[0137] The GICP matching result acquisition unit is used to perform two-dimensional GICP matching on the initial rigid body transformation matrix and the point cloud data of different categories to obtain the GICP matching result.

[0138] Optionally, the GICP matching result acquisition unit includes:

[0139] A two-dimensional point cloud data acquisition subunit is used to convert the point cloud data of different categories into two-dimensional point cloud data on the two-dimensional plane where the vehicle is located; the two-dimensional point cloud data includes: source point cloud data of the vehicle and target point cloud data of the vehicle's perceived environment;

[0140] A point pair set acquisition subunit is used to use a nearest neighbor search algorithm to find the nearest point of each point in the source point cloud data from the target point cloud data in each iteration process to form a corresponding point pair set;

[0141] A residual acquisition subunit is used to calculate, for each corresponding point, the Euclidean distance between the source point and the target point after being transformed by the initial rigid body transformation matrix to obtain a residual;

[0142] A Jacobian equation construction subunit, used for processing the rotation parameter and the translation parameter of the vehicle based on the residual to construct a residual Jacobian equation;

[0143] The GICP matching result acquisition subunit is used to iteratively process the residual Jacobi equation to obtain the GICP matching result.

[0144] Optionally, the GICP matching result includes: an optimized target rigid body transformation matrix,

[0145] The device also includes:

[0146] a transformed point cloud data acquisition module, configured to transform the point cloud data to a position aligned with the vehicle dead reckoning information using the target rigid body transformation matrix to obtain transformed point cloud data when the GICP matching result indicates that the vehicle state has reached an expected state;

[0147] The positioning position acquisition module is used to fuse the transformed point cloud data with the vehicle dead reckoning information to obtain the vehicle positioning position of the vehicle.

[0148] Optionally, the vehicle position information acquisition module includes:

[0149] A particle filter range determination unit, used to determine the particle filter range according to the GICP matching result in the initialization stage of the particle filter;

[0150] a particle state generating unit, configured to perform state prediction for each particle within the particle filtering range according to a motion model, a particle state of the vehicle at a previous moment, and vehicle dead reckoning information, and generate a new particle state, wherein each particle represents a point cloud data in the vehicle dead reckoning information, and the particle state refers to a position state of each point cloud data in the vehicle dead reckoning information;

[0151] a matching degree calculation unit, configured to calculate a matching degree between a new particle state of each particle and the new vehicle dead reckoning information after receiving the new vehicle dead reckoning information;

[0152] A particle weight updating unit, used for updating the particle weight of each particle according to the matching degree to obtain an updated weight of each particle;

[0153] A particle set acquisition unit, used for resampling all particles according to the updated weights to obtain a resampled particle set;

[0154] A system state estimation value acquisition unit, used for processing the resampled particle set to obtain a system state estimation value of the vehicle;

[0155] A vehicle position positioning unit is used to locate and obtain the vehicle position information of the vehicle according to the system state estimation value.

[0156] Optionally, the vehicle position positioning unit includes:

[0157] A confidence calculation subunit, used to calculate the particle filter confidence according to the update weight;

[0158] The vehicle position positioning subunit is used to locate and obtain the vehicle position information of the vehicle according to the system state estimation value when the particle filter confidence is greater than the confidence threshold.

[0159] Optionally, the device further comprises:

[0160] The vehicle position positioning module is used to locate the vehicle position information of the vehicle according to the vehicle track calculation information of the vehicle at the current moment when the particle filter confidence is less than or equal to the confidence threshold.

[0161] The vehicle positioning device provided in the embodiment of the present application obtains the perception detection information and odometer data of the vehicle during the automatic parking process. According to the perception detection information, the point cloud data with boundary semantic element information perceived by the vehicle is determined. The vehicle is subjected to track reckoning processing according to the odometer data to obtain the vehicle track reckoning information corresponding to the vehicle. Two-dimensional GICP matching is performed on the vehicle track reckoning information and the point cloud data to obtain the GICP matching result. When the GICP matching result indicates that the vehicle state has not reached the expected state, the vehicle track reckoning information is filtered and positioned based on the particle filter positioning algorithm to obtain the vehicle position information of the vehicle. The embodiment of the present application completes the semantic GICP matching and particle filter positioning algorithm based on the semantic element information to realize the parking positioning of the vehicle, so as to perform real-time recognition and analysis of the surrounding environment, and can effectively complete the positioning and repositioning functions through environmental information, so as to make the intelligent parking of the vehicle in a complex environment more efficient and reliable.

[0162] In addition, an embodiment of the present application further provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program implements the above-mentioned vehicle positioning method when executed by the processor.

[0163] Figure 4 FIG. 4 is a schematic diagram showing the structure of an electronic device 400 according to an embodiment of the present invention. Figure 4 As shown, the electronic device 400 includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to computer program instructions stored in a read-only memory (ROM) 402 or computer program instructions loaded from a storage unit 408 into a random access memory (RAM) 403. In RAM 403, various programs and data required for the operation of the electronic device 400 can also be stored. CPU 401, ROM 402 and RAM 403 are connected to each other via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0164] Multiple components in the electronic device 400 are connected to the I / O interface 405, including: an input unit 406, such as a keyboard, a mouse, a microphone, etc.; an output unit 407, such as various types of displays, speakers, etc.; a storage unit 408, such as a disk, an optical disk, etc.; and a communication unit 409, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 409 allows the electronic device 400 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0165] The various processes and processing described above may be performed by the processing unit 401. For example, the method of any of the above embodiments may be implemented as a computer software program, which is tangibly contained in a computer-readable medium, such as the storage unit 408. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 400 via the ROM 402 and / or the communication unit 409. When the computer program is loaded into the RAM 403 and executed by the CPU 401, one or more actions in the method described above may be performed.

[0166] The embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, each process of the above-mentioned vehicle positioning method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it is not repeated here. The computer-readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0167] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0168] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0169] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.

[0170] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present application can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0171] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0172] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0173] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0174] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0175] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard drives, ROM, RAM, magnetic disks, or optical disks.

[0176] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A vehicle positioning method, characterized in that: The method comprises: Obtaining the vehicle's perception detection information and odometer data during the automatic parking process; Determining point cloud data with boundary semantic element information perceived by the vehicle according to the perception detection information; Performing dead reckoning processing on the vehicle according to the odometer data to obtain vehicle dead reckoning information corresponding to the vehicle; Performing two-dimensional GICP matching on the vehicle dead reckoning information and the point cloud data to obtain a GICP matching result; When the GICP matching result indicates that the vehicle state has not reached the expected state, the vehicle dead reckoning information is filtered and positioned based on a particle filter positioning algorithm to obtain the vehicle position information of the vehicle.

2. The method according to claim 1, characterized in that The step of determining the point cloud data with boundary semantic element information perceived by the vehicle according to the perception detection information includes: Determining the position information of the perception detection information relative to the vehicle according to the conversion relationship between the inverse perspective mapping image corresponding to the perception detection information and the real position; Converting the perception detection information to a global coordinate system corresponding to the vehicle according to the posture information to obtain conversion coordinates corresponding to the perception detection information; The perception detection information is discretized according to the conversion coordinates to obtain point cloud data with boundary semantic element information perceived by the vehicle.

3. The method according to claim 1, characterized in that The performing two-dimensional GICP matching on the vehicle dead reckoning information and the point cloud data to obtain a GICP matching result includes: Classifying the point cloud data according to the boundary semantic element information to obtain point cloud data of different categories; constructing an initial rigid body transformation matrix based on the vehicle dead reckoning information; Perform two-dimensional GICP matching on the initial rigid body transformation matrix and the point cloud data of different categories to obtain the GICP matching result.

4. The method according to claim 3, characterized in that The performing two-dimensional GICP matching on the initial rigid body transformation matrix and the point cloud data of different categories to obtain the GICP matching result includes: Converting the point cloud data of different categories into two-dimensional point cloud data on the two-dimensional plane where the vehicle is located; the two-dimensional point cloud data includes: source point cloud data of the vehicle and target point cloud data of the vehicle's perceived environment; In each iteration, a nearest neighbor search algorithm is used to find the nearest point of each point in the source point cloud data from the target point cloud data to form a corresponding point pair set; For each corresponding point, the Euclidean distance between the source point and the target point after being transformed by the initial rigid body transformation matrix is ​​calculated to obtain a residual; Processing the rotational parameters and the translational parameters of the vehicle based on the residual to construct a residual Jacobian equation; The residual Jacobi equation is iterated to obtain the GICP matching result.

5. The method according to claim 1, characterized in that The GICP matching results include: optimized target rigid body transformation matrix, After performing two-dimensional GICP matching on the vehicle dead reckoning information and the point cloud data to obtain a GICP matching result, the method further includes: When the GICP matching result indicates that the vehicle state has reached the expected state, transforming the point cloud data to a position aligned with the vehicle dead reckoning information using the target rigid body transformation matrix to obtain transformed point cloud data; The transformed point cloud data and the vehicle dead reckoning information are fused to obtain a vehicle positioning position of the vehicle.

6. The method according to claim 1, characterized in that The filtering and positioning processing of the vehicle dead reckoning information based on the particle filter positioning algorithm to obtain the vehicle position information of the vehicle includes: In the initialization stage of the particle filter, the particle filter range is determined according to the GICP matching result; Within the particle filtering range, a state prediction is performed on each particle according to a motion model, a particle state of the vehicle at a previous moment, and vehicle dead reckoning information, to generate a new particle state, wherein each particle represents a point cloud data in the vehicle dead reckoning information, and the particle state refers to a position state of each point cloud data in the vehicle dead reckoning information; After receiving new vehicle dead reckoning information, calculating a degree of match between a new particle state of each particle and the new vehicle dead reckoning information; According to the matching degree, updating the particle weight of each particle to obtain an updated weight of each particle; Resample all particles according to the updated weights to obtain a resampled particle set; Processing the resampled particle set to obtain an estimated value of the system state of the vehicle; The vehicle position information of the vehicle is located and obtained according to the system state estimation value.

7. The method according to claim 6, characterized in that The step of obtaining the vehicle position information of the vehicle according to the system state estimation value includes: According to the updated weight, the particle filter confidence is calculated; When the particle filter confidence is greater than a confidence threshold, the vehicle position information of the vehicle is located and obtained according to the system state estimation value.

8. The method according to claim 7, characterized in that After calculating the particle filter confidence according to the updated weight, the method further includes: When the particle filter confidence is less than or equal to a confidence threshold, the vehicle position information of the vehicle is located according to the vehicle track dead reckoning information of the vehicle at the current moment.

9. A vehicle positioning device, characterized in that: The device comprises: A vehicle data acquisition module, used to acquire the vehicle's perception detection information and odometer data during the automatic parking process; A point cloud data determination module, used to determine the point cloud data with boundary semantic element information perceived by the vehicle according to the perception detection information; A dead reckoning information acquisition module, configured to perform dead reckoning processing on the vehicle according to the odometer data to obtain vehicle dead reckoning information corresponding to the vehicle; A GICP matching result acquisition module, used for performing two-dimensional GICP matching on the vehicle dead reckoning information and the point cloud data to obtain a GICP matching result; The vehicle position information acquisition module is used to filter and locate the vehicle track calculation information based on the particle filter positioning algorithm to obtain the vehicle position information of the vehicle when the GICP matching result indicates that the vehicle state has not reached the expected state.

10. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the vehicle positioning method according to any one of claims 1 to 8.

11. A readable storage medium, characterized in that: When the instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the vehicle positioning method according to any one of claims 1 to 8.

Images